Retracted: A Rider-Harris Hawks Optimization Algorithm-Based Method for Modular Community Detection in Bipartite Networks

Abstract

Research into the hidden community structure in complex networks has garnered a lot of attention and is now a
hot topic in the multifaceted network domain; this is because this work not only sheds light on the hierarchical
structure of these networks, but it also helps us understand their fundamental functions and, in turn, how to
recommend information to users. In a bipartite network, each set of nodes is distinct from the other, and no edges
connect the vertices. This kind of network is known as a multidimensional network. Despite much research on
community identification in one-mode networks, bipartite network community detection remains unexplored.
Here, we provide a new method for community discovery in bipartite networks using node similarity—the Rider-
Harris Hawks Optimisation (RHHO) algorithm. We devised the suggested RHHO by integrating the RO and HHO
algorithms, which stand for Rider Optimisation and Harris Hawks Optimisation, respectively. Additionally, a
novel assessment measure called the h-Tversky Index (h-TI) is put forth for the purpose of calculating node
similarity and fitness, with modularity as a key consideration. By breaking the network into smaller, more
manageable pieces, we can test how well the suggested community identification works. This is what modularity
is all about. Over the citation networks (cit-HepPh and cit-HepTh) and bipartite networks (Movie Lens 100 K and
American Revolution datasets), the suggested method was quantitatively evaluated and compared in terms of
robustness and modularity. After 250 simulation tests, the results were analysed. The proposed method
outperformed state-of-the-art approaches, including h-index-based link prediction, multiagent genetic algorithms
(MAGA), genetic algorithms (GA), memetic algorithm for community detection in bipartite networks
(MATMCD-BN), and HHO, according to experimental results. It also showed a maximal fitness of 0.74353 and
maximal modularity of 0.77433